다양한 분류 알고리즘

본 내용은 혼자 공부하는 머신러닝 코드에 출처가 있음을 밝힙니다.

로지스틱 회귀

# 데이터를 불러오는 코드
import pandas as pd
fish = pd.read_csv('https://bit.ly/fish_csv_data')
fish.head()

print(pd.unique(fish['Species']))

['Bream' 'Roach' 'Whitefish' 'Parkki' 'Perch' 'Pike' 'Smelt']

# 생선 input과 target을 나눈다. 
fish_input = fish[['Weight','Length','Diagonal','Height','Width']].to_numpy()
fish_target = fish['Species'].to_numpy()

fish_input.shape

(159, 5)

# train test data split
from sklearn.model_selection import train_test_split
train_input, test_input, train_target, test_target = train_test_split(fish_input, fish_target, random_state = 42)

# standard scaler 사용하여 번환
from sklearn.preprocessing import StandardScaler
ss = StandardScaler()
ss.fit(train_input)
train_scaled = ss.transform(train_input )
test_scaled = ss.transform(test_input)

데이터의 크기가 feature에 따라 다르기 때문에 standardscaler로 크기 변환을 해준다. 이는 거리를 재는 머신러닝에서 특히 유용하게 쓰인다.

# KNN으로 클래스 분류해보기
from sklearn.neighbors import KNeighborsClassifier
kn = KNeighborsClassifier(n_neighbors=3)
kn.fit(train_scaled, train_target)
print(kn.score(train_scaled, train_target))
print(kn.score(test_scaled, test_target))
print(kn.classes_)

0.8907563025210085
0.85
['Bream' 'Parkki' 'Perch' 'Pike' 'Roach' 'Smelt' 'Whitefish']

print(kn.predict(test_scaled[:10]))

['Perch' 'Smelt' 'Pike' 'Perch' 'Perch' 'Bream' 'Smelt' 'Roach' 'Perch'
 'Pike']

# 어떻게 예측한 것일까?(확률)
proba = kn.predict_proba(test_scaled[:5])
print(np.round(proba, decimals = 4)) # 소수 4번째 자리에서 표기

[[0.     0.     1.     0.     0.     0.     0.    ]
 [0.     0.     0.     0.     0.     1.     0.    ]
 [0.     0.     0.     1.     0.     0.     0.    ]
 [0.     0.     0.6667 0.     0.3333 0.     0.    ]
 [0.     0.     0.6667 0.     0.3333 0.     0.    ]]

distance, indexes = kn.kneighbors(test_scaled[3:4])
print(train_target[indexes])

[['Roach' 'Perch' 'Perch']]

## 시그모이드 함수 그래프 그리기
import numpy as np
import matplotlib.pyplot as plt
plt.style.use(['seaborn'])
z = np.arange(-5, 5, 0.1)
phi = 1 / (1 + np.exp(-z))
plt.plot(z, phi)
plt.xlabel('Z')
plt.ylabel('PHI')
plt.show()
# 사이킷런은 0.5일 때, 음성 클래스로 판단한다.

## 로지스틱 회귀로 이진분류 수행하기
char_arr = np.array(['A','B','C','D','E'])
print(char_arr[[True, False,True, False,False]])

['A' 'C']

bream_smelt_indexes = (train_target == 'Bream') | (train_target == 'Smelt')
train_bream_smelt = train_scaled[bream_smelt_indexes]
target_bream_smelt = train_target[bream_smelt_indexes]

from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
lr.fit(train_bream_smelt, target_bream_smelt)
lr.predict(train_bream_smelt)

array(['Bream', 'Smelt', 'Bream', 'Bream', 'Bream', 'Smelt', 'Bream',
       'Bream', 'Bream', 'Bream', 'Bream', 'Bream', 'Bream', 'Smelt',
       'Bream', 'Smelt', 'Smelt', 'Bream', 'Bream', 'Bream', 'Bream',
       'Bream', 'Bream', 'Bream', 'Bream', 'Smelt', 'Bream', 'Smelt',
       'Smelt', 'Bream', 'Smelt', 'Bream', 'Bream'], dtype=object)

lr.predict_proba(train_bream_smelt[:5])

array([[0.99759855, 0.00240145],
       [0.02735183, 0.97264817],
       [0.99486072, 0.00513928],
       [0.98584202, 0.01415798],
       [0.99767269, 0.00232731]])

print(lr.classes_)
print(lr.coef_, lr.intercept_)

['Bream' 'Smelt']
[[-0.4037798  -0.57620209 -0.66280298 -1.01290277 -0.73168947]] [-2.16155132]

lr = LogisticRegression(C=20, max_iter = 1000)
lr.fit(train_scaled, train_target)
print(lr.score(train_scaled, train_target))
print(lr.score(test_scaled,test_target))

0.9327731092436975
0.925

proba = lr.predict_proba(test_scaled[:5])
print(np.round(proba, decimals =3))
print(lr.classes_)

[[0.    0.014 0.841 0.    0.136 0.007 0.003]
 [0.    0.003 0.044 0.    0.007 0.946 0.   ]
 [0.    0.    0.034 0.935 0.015 0.016 0.   ]
 [0.011 0.034 0.306 0.007 0.567 0.    0.076]
 [0.    0.    0.904 0.002 0.089 0.002 0.001]]
['Bream' 'Parkki' 'Perch' 'Pike' 'Roach' 'Smelt' 'Whitefish']

손실함수

fish

from sklearn.model_selection import train_test_split
train_input, test_input, train_target, test_target = train_test_split(fish_input, fish_target, random_state = 42)

from sklearn.preprocessing import StandardScaler
ss = StandardScaler()
ss.fit(train_input)
train_scaled = ss.transform(train_input)
test_scaled = ss.transform(test_input)

# 확률적 경사 하강법
from sklearn.linear_model import SGDClassifier
sc = SGDClassifier(loss = 'log', max_iter = 10, random_state = 42)
sc.fit(train_scaled, train_target)
print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))

0.773109243697479
0.775


/Users/junhyeoungson/opt/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_stochastic_gradient.py:573: ConvergenceWarning: Maximum number of iteration reached before convergence. Consider increasing max_iter to improve the fit.
  ConvergenceWarning)

# 모델을 계속 훈련
sc.partial_fit(train_scaled, train_target)
print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled,test_target))

0.8319327731092437
0.8

# 모델 정확도 그래프

sc = SGDClassifier(loss= 'log', random_state = 42)
train_score = []
test_score = []
classes = np.unique(train_target)
for _ in range(0, 300):
    sc.partial_fit(train_scaled, train_target, classes = classes)
    train_score.append(sc.score(train_scaled, train_target))
    test_score.append(sc.score(test_scaled, test_target))

plt.plot(train_score)
plt.plot(test_score)

[<matplotlib.lines.Line2D at 0x7fcb9bfd4090>]

sc = SGDClassifier(loss= 'log', random_state = 42,tol = None, max_iter = 100) # tol - 향상될 최솟값 지정
sc.fit(train_scaled, train_target)
print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))

0.957983193277311
0.925

# 힌지 사용
sc = SGDClassifier(loss= 'hinge', random_state = 42,tol = None, max_iter = 100) # tol - 향상될 최솟값 지정
sc.fit(train_scaled, train_target)
print(sc.score(train_scaled, train_target))
print(sc.score(test_scaled, test_target))

0.9495798319327731
0.925